Team 5 Gaming Data Notebook

Team Members: Fucheng Yao, Huaiping Wang, Limei Huang, Eman Nagib, Kwangwoo Kim

Gaming Data Overview and Backgrond Knowledge

1.Brief description and data soure

The gaming data set has information of video games with sales greater than 100,000 copies since the year 1980 to 2016. This gaming data set is downloaded from Kaggle.com.

Mobile, PC/Mac(computer), social/online, and console are common gaming platforms. Console here refers to a computer device that outputs a video signal or visual image to display a video game that one or more people can play. Some popular consoles are PlayStation 4 Pro, Xbox One X, Nintendo Switch. Our dataset is about videogames played on console only.

---

2.Data Cleaning

• We \(\color{red}{\text{removed}}\) all NULL values, Unknown values, and data in 2017 or later.
• We \(\color{red}{\text{converted}}\) Year from String to Numeric.
• We \(\color{red}{\text{added}}\) a new column called Portable based on if the gaming console is portable or not.
• We have 16187 rows and 12 variables after cleaning the data and adding the Portable variable.

options(stringsAsFactors = FALSE)
Portable <- function(df) {
 len <- length(df$Platform)
 new_vec <- vector(mode = "numeric", length = len)
 protvec <- c("DS", "GB", "3DS", "GBA")
 for (i in 1:len)  {
   if (df$Platform[i] %in% protvec) {
     new_vec[i] <- 1
   } else {
     new_vec[i] <- 0
   }
 }
 return(new_vec)
}

---

3.Variables

Variable	Description
Rank	Ranking of overall sales
Name	The games name
Platform	Platform of the games release (i.e. PC,PS4, etc.)
Year	Year of the game’s release
Genre	Genre - Genre of the game
Publisher	Publisher of the game
NA_Sales	Sales in North America (in millions)
EU_Sales	Sales in Europe (in millions)
JP_Sales	Sales in Japan (in millions)
Other_Sales	Sales in the rest of the world (in millions)
Global_Sales	Total worldwide sales
Portable	If the gaming console is portable (1=yes,0=no)

---

4.Glimpse of Data

a.

Classes ‘spec_tbl_df’, ‘tbl_df’, ‘tbl’ and 'data.frame':    16187 obs. of  12 variables:
 $ Rank        : num  3825 1679 1879 1711 638 ...
 $ Name        : chr  "Seaman" "NFL 2K" "NFL 2K1" "Shenmue" ...
 $ Platform    : chr  "DC" "DC" "DC" "DC" ...
 $ Year        : num  1999 1999 2000 1999 1998 ...
 $ Genre       : chr  "Simulation" "Sports" "Sports" "Adventure" ...
 $ Publisher   : chr  "Sega" "Sega" "Sega" "Sega" ...
 $ NA_Sales    : num  0 1.12 1.02 0.52 1.26 1.1 0.41 0 0 0 ...
 $ EU_Sales    : num  0 0.05 0.05 0.24 0.61 0.51 0.23 0 0 0 ...
 $ JP_Sales    : num  0.52 0 0 0.38 0.46 0.12 0.47 1.01 0.54 0.62 ...
 $ Other_Sales : num  0 0.02 0.02 0.04 0.08 0.08 0.03 0 0 0 ...
 $ GLobal_Sales: num  0.52 1.2 1.09 1.18 2.42 1.81 1.14 1.01 0.54 0.62 ...
 $ Portable    : num  0 0 0 0 0 0 0 1 1 1 ...

b. 31 unique platforms:

DC, DS, GB, GC, PC, PS, WS, XB, 3DS, GBA, GEN, N64, NES, PS2, PS3, PS4, PSP, PSV, SAT, SCD, Wii, 2600, SNES, WiiU, X360, XOne, NG, 3DO, TG16, GG, PCFX

c. 12 unique genres:

Simulation, Sports, Adventure, Platform, Racing, Action, Misc, Role-Playing, Puzzle, Fighting, Strategy, Shooter

d. 575 unique publishers:

Top 10 Publishers based on frequency: Electronic Arts, Activision, Namco Bandai Games, Ubisoft, Konami Digital Entertainment, THQ, Nintendo, Sony Computer Entertainment, Sega, Take-Two Interactive

---

Problem Set: Where and how to invest in the gaming industry

• Global trends and anlysis
• Regional trends and analysis
• Investment options and recommendations

---

1.Global trends and analysis

1.a What is the top platform each year?

• Once a platform clicks in the market, it goes on to rule for a few years.
• The Playstation platform was popular for nearly 20 years.

top_platforms <- games %>%
             group_by(Year, Platform) %>%
             summarize(Revenue = sum(GLobal_Sales)) %>%
             arrange(desc(Revenue)) %>%
             top_n(1)

ggplot(top_platforms, aes(Year, Revenue, fill = Platform)) + 
  geom_bar(stat = "identity") +
  ggtitle("Top Platform by Revenue each year") +
  theme(legend.position = "top") + 
  scale_fill_manual(values = mycolors)+
  scale_x_continuous(breaks = seq(min(top_platforms$Year),max(top_platforms$Year),5))+
  theme(axis.text.x = element_text(angle = 90))

1.b Top 10 best selling publishers over time?

• Top 3 publishers are Nintendo, EA, and Activision.

top10 <- games %>%select(Publisher,GLobal_Sales)%>%group_by(Publisher)%>%summarise(GLobal_Sales=sum(GLobal_Sales))%>%arrange(desc(GLobal_Sales))%>%head(10)


p <- ggplot(top10, aes(x=reorder(Publisher,-GLobal_Sales),y=GLobal_Sales,,label=round(GLobal_Sales,2))) +
 stat_summary(fun.y=sum, geom="bar",position=position_dodge(1),width=0.8,show.legend = F,,col="black",fill="skyblue")+
labs(title = "Top 10 best selling publishers",caption = "Sales in Million")+
   geom_text(col="black",size=4,vjust=-1)+
  ylab("Global Sales")+xlab("Publisher")+
  scale_x_discrete(labels = function(x) str_wrap(x, width =1))

p

1.c Best selling platforms for the TOP 3 publishers

• Top 3 platforms for each publisher:
• Nintendo: Wii, GB, DS
• EA: X360, PS3, PS2
• Activision: X360, PS3, PS2

top_three_publisher <- subset(games,Publisher %in% c("Nintendo","Electronic Arts","Activision"))
top_three_publisher$Portable <- factor(top_three_publisher$Portable)
top_three_publisher_platform <- top_three_publisher %>%select(Publisher,Platform,Portable,GLobal_Sales)%>%group_by(Publisher,Platform,Portable)%>%summarise(GLobal_Sales=sum(GLobal_Sales))%>%arrange(desc(Publisher))
top_new <- top_three_publisher_platform%>%group_by(Publisher)%>%top_n(3)

top_new$Publisher_f = factor(top_new$Publisher, levels=c("Nintendo","Electronic Arts","Activision"))
ggplot(top_new, aes(x=Platform, y=GLobal_Sales, color=Publisher,shape=Portable)) +
 geom_point(size=3)+
 scale_color_discrete(breaks=c("Nintendo","Electronic Arts","Activision")) +
 labs(title="     Top 3 platforms for the top 3 publishers",caption="Sales in Million") +
 geom_segment(aes(x=Platform,xend=Platform, y=0, yend=GLobal_Sales))+
 geom_text(aes(label=GLobal_Sales), hjust = -0.3, size = 2.6,fontface = "bold",color='black') +
 theme( plot.title = element_text(size=17,hjust=-0.4)) +
   facet_wrap(~ Publisher_f, nrow = 5, scales = 'free', strip.position = 'right')+
 ylim(0, max(top_new$GLobal_Sales + 10))+ylab("Global Sales")+
 coord_flip()

1.d Best selling genre for the TOP 3 publishers

• Top 3 genres for each publisher:
• Nintendo: Sports, Role-playing, Platform
• EA: sports, shooter, racing
• Activision: sports, shooter, action

top_three_publisher <- subset(games,Publisher %in% c("Nintendo","Electronic Arts","Activision"))
top_three_publisher$Portable <- factor(top_three_publisher$Portable)
top_three_publisher_platform <- top_three_publisher %>%
select(Publisher,Genre,Portable,GLobal_Sales)%>%
group_by(Publisher,Genre,Portable)%>%
summarise(GLobal_Sales=sum(GLobal_Sales))%>%
arrange(desc(GLobal_Sales))
top_new <- top_three_publisher_platform%>%group_by(Publisher)%>%top_n(3)

top_new$Publisher_g = factor(top_new$Publisher, levels=c("Nintendo","Electronic Arts","Activision"))
ggplot(top_new, aes(x=Genre, y=GLobal_Sales, color=Publisher,shape=Portable)) +
  geom_point(size=3)+
scale_color_discrete(breaks=c("Nintendo","Electronic Arts","Activision"))+
labs(title="The top 3 genre for the top 3 publishers",caption="Sales in Million") +
geom_segment(aes(x=Genre,xend=Genre, y=0, yend=GLobal_Sales))+
geom_text(aes(label=GLobal_Sales), hjust = -0.3, size = 2.6,fontface = "bold",color='black') +
theme( plot.title = element_text(size=17,hjust=-0.5)) +
  facet_wrap(~ Publisher_g, nrow = 5, scales = 'free', strip.position = 'right')+
ylim(0, max(top_new$GLobal_Sales + 10))+ylab("Global Sales")+
coord_flip()

1.e Genre frequency over the years and Platform popularity over the years

• Here, the popularity is defined as the number of games published on each platform. In the early 1980s, there were limited amount of platforms in the market (2600, NES), but later on, more and more platforms emerged. It is also interesting to see that different platform tends to dominate the market in different periods, and there is a trend to shift from less portable to portable ones.

Years	Most Popular Platform	Most Profitable Platform
1980 - 1982	2600	2600
1983	2600	NES
1984 - 1988	NES	NES
1989	GB	GB
1990	NES	SNES
1991 - 1994	SNES	SNES
1995 - 2000	PS	PS
2001 - 2005	PS2	PS2
2006	PS2	Will
2007 - 2010	DS	Will
2011 - 2014	PS3	PS3
2015 - 2016	PS4	PS4

Ignoring unknown aesthetics: frame

2. Regional trends and analysis

2.a Top 3 best selling platform by regions

• We studied the regional trends by looking at the top platforms for each region as shown from the bar chart.

games <- games[!(games$Year %in% c("N/A", "2017", "2020")),]
games <- games %>% gather(Region, Revenue, 7:10) 
games$Region <- factor(games$Region)

mytheme_1 <- function() {
  
 return(theme(axis.text.x = element_text(angle = 90, size = 10, vjust = 0.4), plot.title = element_text(size = 15, vjust = 2),axis.title.x = element_text(size = 12, vjust = -0.35)))
  
}

mytheme_2 <- function() {
  
 return(theme(axis.text.x = element_text(size = 10, vjust = 0.4), plot.title = element_text(size = 15, vjust = 2),axis.title.x = element_text(size = 12, vjust = -0.35)))
  
}

mycolors <- c("#771C19", "#AA3929", "#8E9CA3", "#556670", "#000000", "#E25033", "#F27314", "#F8A31B", "#E2C59F", "#B6C5CC","#771C19", "#AA3929", "#8E9CA3", "#556670", "#000000", "#E25033", "#F27314", "#F8A31B", "#E2C59F", "#B6C5CC","#771C19", "#AA3929", "#8E9CA3", "#556670", "#000000", "#E25033", "#F27314", "#F8A31B", "#E2C59F", "#B6C5CC","#8E9CA3")


top_platform_region <- games %>%
             group_by(Region, Platform) %>%
             summarize(Revenue = sum(Revenue)) %>%
             arrange(desc(Revenue)) %>%
             top_n(3)

Selecting by Revenue

ted <- ggplot(top_platform_region, aes(Region, Revenue, fill = Platform)) + 
  geom_bar(position = "dodge", stat = "identity")  +
  ggtitle("Top 3 best selling platform by regions") +
  ylab("Revenue in Millions") +
  xlab("Region") +
  mytheme_2() +
  theme(legend.position = "top") + 
  scale_fill_manual(values = c("#8E9CA3","#F8A31B", "#AA3929", "#E25033", "#E2C59F", "#556670"))
ggplotly(ted)

2.b Top 3 best selling publisher by regions

• We studied the regional trends by looking at the top 3 best selling publishers for each region.

top_genres_region <- games %>%
             group_by(Region, Publisher) %>%
             summarize(Revenue = sum(Revenue)) %>%
             arrange(desc(Revenue)) %>%
             top_n(3)

ted2 <- ggplot(top_genres_region, aes(Region, Revenue, fill = Publisher)) + 
  geom_bar(position = "dodge", stat = "identity")  +
  ggtitle("Top 3 best selling publisher by region") +
  ylab("Sales in Millions") +
  xlab("Region") +
  mytheme_2() +
  theme(legend.position = "top")

ggplotly(ted2)

2.c Best selling genre for particular regions

*This heat map identfies the top selling genres for each region by displaying a deeper color (purple) for genres with high revenues.

year_genre <- games %>% 
                group_by(Year, Genre, Region) %>% 
                  summarise(TotalRevenue = sum(Revenue)) 
                

ggplot(year_genre, aes(Year, Genre, fill = TotalRevenue)) +
    geom_tile(color = "white") +
    ggtitle("                       Best selling genre for particular regions") + 
    facet_wrap(vars(Region), ncol = 4) +
theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank())+
  scale_color_gradient(low="pink", high= "purple")+
    scale_fill_gradient(low="pink", high= "purple")

2.d Top 3 best selling genre by region

• We studied the regional trends by looking at the top 3 best selling genres for each region.

top_genres_region <- games %>%
             group_by(Region, Genre) %>%
             summarize(Revenue = sum(Revenue)) %>%
             arrange(desc(Revenue)) %>%
             top_n(3)

ted2 <- ggplot(top_genres_region, aes(Region, Revenue, fill = Genre)) + 
  geom_bar(position = "dodge", stat = "identity")  +
  ggtitle("Top 3 best selling genre by region") +
  ylab("Revenue in Millions") +
  xlab("Region") +
  mytheme_2() +
  theme(legend.position = "top")

ggplotly(ted2)

3 Investment options and recommendations

3.a Global sales proportion by region

• For the most part we see that the North America accounts for the highest proportion of global sales. We can also see that European sales are on an incline and actually surpass North american sales by the year 2015-2016. Even though Japan’s proportion of Global sales seem to be declining in the past years, for the most recent years it seems to be steadily inclining.

games <- read_csv("games.csv") 

df_trial <- data_frame(sort(games$Year), NA_Sales = games$NA_Sales, games$EU_Sales,
                      games$JP_Sales, games$Other_Sales, games$GLobal_Sales)
games3 <- games %>%
 select(Year, NA_Sales, EU_Sales, JP_Sales, Other_Sales,GLobal_Sales) %>%
 group_by(Year) %>%
 summarise(NA_sales_prop = sum(NA_Sales)/sum(GLobal_Sales),
           EU_sales_prop = sum(EU_Sales)/sum(GLobal_Sales),
           JP_sales_prop = sum(JP_Sales)/sum(GLobal_Sales),
           Other_sales_prop = sum(Other_Sales)/sum(GLobal_Sales),
           Global_sales_prop = sum(GLobal_Sales)/sum(GLobal_Sales))
mycolors <- c("#771C19", "#AA3929", "#8E9CA3", "#556670", "#000000", "#E25033", "#F27314", "#F8A31B", "#E2C59F", "#B6C5CC")
regions <- c( "darkgreen" = "North America", "blue" = "Europe" ,  "sienna" = "Japan" , "orange" = "Other Regions", "black" = "Global")

ggplot(games3,aes(x=Year))+
  geom_line(aes(y = NA_sales_prop ,color = "#771C19"))+
  geom_line(aes(y = EU_sales_prop , color = "#B6C5CC"))+
  geom_line(aes(y = JP_sales_prop, color = "#E25033" ))+
  geom_line(aes(y = Other_sales_prop, color = "#E2C59F"))+
  geom_line(aes(y = Global_sales_prop, color = "black"))+
  labs(title = "Sales per region from 1980-2016", y = "Percentage of global sales")+
  scale_color_manual(name = "Regions", values = c( "darkgreen", "blue", "sienna" , "orange", "black"),labels=c("North America", "Europe" , "Japan" , "Other Regions", "Global"))

3.b Summary

Option	North America	Europe	Japan
Publisher	Nintendo	Nintendo	Nintendo
Platform	X360	PS3	DS
Genre	Action	Action	Role-Playing

Appendix

Data Exploration.

ggplot(games, aes(x=Year, fill=..count..)) +
    geom_bar()+
    scale_color_gradient(low="#771C19", high= "#F27314")+
    scale_fill_gradient(low="#771C19", high= "#F27314")+
    labs(title="Number of Games Released every Year", x= "Year", 
         y= "Total Number of Games")+
    geom_text(stat='count',aes(label=..count..), hjust=-0.1,color="black", size=2.5)+
    scale_x_continuous(breaks = 1980:2016) + theme_minimal()+
  coord_flip()

Data Exploration 2

Data Exploration 3

ok <- games %>% select(Year,GLobal_Sales,Genre)%>%group_by(Year,Genre)%>%
  summarise(Total_sales=sum(GLobal_Sales)) 
ok1 <- arrange(ok, desc(Year))

plot_ly(ok1, x = ~Total_sales, y = ~Genre, z = ~Year) %>% layout( title ="Sales by genre from 1980 - 2016") %>%
  add_markers(color = ~Genre, size = 0.5)